Enter opens the highlighted result. Esc closes.

Qwen3.5-35B-A3B vs Qwen3.8 Max (0902)

Qwen3.8 Max (0902) leads Qwen3.5-35B-A3B on the Artificial Analysis Intelligence Index, 45.4 to 19.3, and Qwen3.5-35B-A3B costs 4.6x less per output token ($1.30 vs $6.00 per million). Updated October 22, 2026; 41 benchmarks have a result for both.

Replace Qwen3.5-35B-A3B Replace Qwen3.8 Max (0902)
Qwen3.5-35B-A3BQwen
Coding
No data
Overall
53.5
Output $/M
$1.30
0
Qwen3.8 Max (0902)Qwen
Coding
62.4
Overall
68.9
Output $/M
$6.00
41
vs boards won of 41 shared

Qwen3.5-35B-A3B or Qwen3.8 Max (0902): which to pick

Only Qwen3.8 Max (0902) has a coding score, 62.4 on the Vibe Coding Index; Qwen3.5-35B-A3B has none, so the two cannot be ranked on coding.

Pick Qwen3.5-35B-A3B for output tokens 4.6x cheaper ($1.30 vs $6.00 per million).

Model AQwen3.5-35B-A3BQwenNot benchmarked1 independent benchmarkNo cost per solved taskModel BQwen3.8 Max (0902)Qwen62.4/1005 independent benchmarks$1.32 per solved task
Shared results
41
Qwen3.5-35B-A3B only
5
Qwen3.8 Max (0902) only
73
Like-for-like categories
6 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
Qwen3.5-35B-A3B not benchmarked, Qwen3.8 Max (0902) 62.4 from 5 benchmarks.
Widest category
Coding, 39.9 points apart on average over 5 shared benchmarks.
Widest single gap
LMArena WebDev: React: Qwen3.5-35B-A3B 1238 against Qwen3.8 Max (0902) 1692, 60.3 points apart on a 0 to 100 scale.
Output price
Qwen3.5-35B-A3B $1.30 against Qwen3.8 Max (0902) $6.00 per million output tokens, 4.6x apart.
Context window
Qwen3.5-35B-A3B takes 262K tokens, Qwen3.8 Max (0902) 1M.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryQwen3.5-35B-A3BBarsQwen3.8 Max (0902)Ties
Coding05
Intelligence02
Reasoning06
Math04
Knowledge04
General020
All shared boards041
CodingLike-for-likeSix of these boards feed the Vibe Coding Index05
Coding benchmarks for Qwen3.5-35B-A3B and Qwen3.8 Max (0902)
BenchmarkQwen3.5-35B-A3BQwen3.8 Max (0902)Leader
LiveBench codingNot measured60.3%Only Qwen3.8 Max (0902) measured
LMArena WebDev12541674Qwen3.8 Max (0902)
AA Coding IndexNot measured76.2Only Qwen3.8 Max (0902) measured
LiveCodeBench (Vals AI)Not measured87.8%Only Qwen3.8 Max (0902) measured
SWE-bench Verified (Vals AI)Not measured85.6%Only Qwen3.8 Max (0902) measured
Terminal-Bench 4 (Vals AI)Not measured34.3%Only Qwen3.8 Max (0902) measured
Vibe Code Bench (Vals AI)Not measured64.7%Only Qwen3.8 Max (0902) measured
Code migration (Vals AI)Not measured24.0%Only Qwen3.8 Max (0902) measured
ProgramBench (Vals AI)Not measured0.0%Only Qwen3.8 Max (0902) measured
IOI (Vals AI)Not measured68.9%Only Qwen3.8 Max (0902) measured
Terminal-Bench 2.1 (Vals AI)Not measured67.4%Only Qwen3.8 Max (0902) measured
SWE-bench Multimodal (Vals AI)Not measured42.5%Only Qwen3.8 Max (0902) measured
Vibe Code Bench 1-100 (Vals AI)Not measured12.8%Only Qwen3.8 Max (0902) measured
SkillsBench (Vals AI)Not measured42.0%Only Qwen3.8 Max (0902) measured
CyberBench (Vals AI)Not measured28.6%Only Qwen3.8 Max (0902) measured
SciCodeNot measured53.2%Only Qwen3.8 Max (0902) measured
Terminal-Bench Hard (Artificial Analysis run)26.5%Not measuredOnly Qwen3.5-35B-A3B measured
FrontierSWENot measured17.8%Only Qwen3.8 Max (0902) measured
ALE-Bench221.8Not measuredOnly Qwen3.5-35B-A3B measured
LMArena Text: coding14101502Qwen3.8 Max (0902)
LMArena Text: software and IT14151498Qwen3.8 Max (0902)
LMArena WebDev: React12381692Qwen3.8 Max (0902)
LMArena WebDev: HTML13011649Qwen3.8 Max (0902)
LMArena WebDev: from an imageNot measured1637Only Qwen3.8 Max (0902) measured
Neither model has a result on the other 19 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for Qwen3.5-35B-A3B and Qwen3.8 Max (0902)
BenchmarkQwen3.5-35B-A3BQwen3.8 Max (0902)Leader
AA Intelligence Index19.345.4Qwen3.8 Max (0902)
Epoch Capabilities Index142.5157.4Qwen3.8 Max (0902)
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index20 boards, none shared
Agentic work benchmarks for Qwen3.5-35B-A3B and Qwen3.8 Max (0902)
BenchmarkQwen3.5-35B-A3BQwen3.8 Max (0902)Leader
AA Agentic IndexNot measured56.0Only Qwen3.8 Max (0902) measured
Design Arena: agents building web appsNot measured1300Only Qwen3.8 Max (0902) measured
Design Arena: agents building full-stack appsNot measured1305Only Qwen3.8 Max (0902) measured
Design Arena: agents building mobile appsNot measured1224Only Qwen3.8 Max (0902) measured
Design Arena: agents building Android appsNot measured1190Only Qwen3.8 Max (0902) measured
Design Arena: agents building Godot gamesNot measured1234Only Qwen3.8 Max (0902) measured
Design Arena: PowerPoint slidesNot measured1232Only Qwen3.8 Max (0902) measured
Design Arena: spreadsheetsNot measured993Only Qwen3.8 Max (0902) measured
LiveBench researchNot measured69.9%Only Qwen3.8 Max (0902) measured
LiveBench browserNot measured32.0%Only Qwen3.8 Max (0902) measured
LiveBench long agenticNot measured42.3%Only Qwen3.8 Max (0902) measured
Finance Agent (Vals AI)Not measured50.6%Only Qwen3.8 Max (0902) measured
Tax Agent Bench (Vals AI)Not measured32.1%Only Qwen3.8 Max (0902) measured
Legal Research Bench (Vals AI)Not measured47.6%Only Qwen3.8 Max (0902) measured
Excel Modeling Benchmark (Vals AI)Not measured60.1%Only Qwen3.8 Max (0902) measured
Terminal-Bench-Science (Vals AI)Not measured12.9%Only Qwen3.8 Max (0902) measured
τ²-Bench Telecom89.2%Not measuredOnly Qwen3.5-35B-A3B measured
APEX-AgentsNot measured63.3%Only Qwen3.8 Max (0902) measured
Vending-Bench 2$463Not measuredOnly Qwen3.5-35B-A3B measured
LMArena SearchNot measured1261Only Qwen3.8 Max (0902) measured
Neither model has a result on the other 22 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index06
Reasoning benchmarks for Qwen3.5-35B-A3B and Qwen3.8 Max (0902)
BenchmarkQwen3.5-35B-A3BQwen3.8 Max (0902)Leader
GPQA DiamondNot measured93.7%Only Qwen3.8 Max (0902) measured
LiveBench reasoningNot measured88.2%Only Qwen3.8 Max (0902) measured
MysteryMechanism (Vals AI)Not measured23.9%Only Qwen3.8 Max (0902) measured
MMMU Pro (Vals AI)Not measured88.0%Only Qwen3.8 Max (0902) measured
GPQA Diamond (Artificial Analysis run)84.5%92.8%Qwen3.8 Max (0902)
Humanity's Last Exam (Artificial Analysis run)21.0%43.1%Qwen3.8 Max (0902)
GPQA Diamond (Epoch AI run)83.5%92.7%Qwen3.8 Max (0902)
Chess puzzlesNot measured46.0%Only Qwen3.8 Max (0902) measured
Mystery game puzzlesNot measured38.0%Only Qwen3.8 Max (0902) measured
LMArena Text: hard prompts14001496Qwen3.8 Max (0902)
LMArena Text: expert14081507Qwen3.8 Max (0902)
LMArena Text: hard prompts in English14111503Qwen3.8 Max (0902)
LMArena Vision: diagramsNot measured1323Only Qwen3.8 Max (0902) measured
LMArena Vision: homeworkNot measured1336Only Qwen3.8 Max (0902) measured
Neither model has a result on the other 9 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index04
Math benchmarks for Qwen3.5-35B-A3B and Qwen3.8 Max (0902)
BenchmarkQwen3.5-35B-A3BQwen3.8 Max (0902)Leader
OTIS Mock AIME70.0%100.0%Qwen3.8 Max (0902)
FrontierMath, tiers 1 to 3Not measured74.7%Only Qwen3.8 Max (0902) measured
FrontierMath, tier 4Not measured46.3%Only Qwen3.8 Max (0902) measured
ProofBenchNot measured58.0%Only Qwen3.8 Max (0902) measured
LMArena Text: math14041501Qwen3.8 Max (0902)
LMArena Text: mathematical work14101491Qwen3.8 Max (0902)
MathArena14.7%51.6%Qwen3.8 Max (0902)
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index04
Knowledge benchmarks for Qwen3.5-35B-A3B and Qwen3.8 Max (0902)
BenchmarkQwen3.5-35B-A3BQwen3.8 Max (0902)Leader
MMLU ProNot measured88.6%Only Qwen3.8 Max (0902) measured
LegalBench (Vals AI)Not measured83.6%Only Qwen3.8 Max (0902) measured
CorpFin (Vals AI)Not measured65.8%Only Qwen3.8 Max (0902) measured
TaxEval (Vals AI)Not measured75.5%Only Qwen3.8 Max (0902) measured
MortgageTax (Vals AI)Not measured64.0%Only Qwen3.8 Max (0902) measured
MedCode (Vals AI)Not measured40.7%Only Qwen3.8 Max (0902) measured
SimpleQA VerifiedNot measured47.3%Only Qwen3.8 Max (0902) measured
LMArena Text: science14131512Qwen3.8 Max (0902)
LMArena Text: medicine and healthcare14211503Qwen3.8 Max (0902)
LMArena Text: legal and government13991492Qwen3.8 Max (0902)
LMArena Text: business and finance13941480Qwen3.8 Max (0902)
Neither model has a result on the other 9 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
Web design benchmarks for Qwen3.5-35B-A3B and Qwen3.8 Max (0902)
BenchmarkQwen3.5-35B-A3BQwen3.8 Max (0902)Leader
Design Arena: websitesNot measured1275Only Qwen3.8 Max (0902) measured
Design Arena: UI componentsNot measured1336Only Qwen3.8 Max (0902) measured
Design Arena: game devNot measured1312Only Qwen3.8 Max (0902) measured
Design Arena: data vizNot measured1282Only Qwen3.8 Max (0902) measured
Design Arena: 3DNot measured1336Only Qwen3.8 Max (0902) measured
Design Arena: all code categoriesNot measured1292Only Qwen3.8 Max (0902) measured
Design Arena: ASCII artNot measured1200Only Qwen3.8 Max (0902) measured
LiveBench designNot measured78.6%Only Qwen3.8 Max (0902) measured
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index020
General benchmarks for Qwen3.5-35B-A3B and Qwen3.8 Max (0902)
BenchmarkQwen3.5-35B-A3BQwen3.8 Max (0902)Leader
LiveBench data analysisNot measured78.4%Only Qwen3.8 Max (0902) measured
Public Benefits Bench (Vals AI)Not measured67.1%Only Qwen3.8 Max (0902) measured
SAGE (Vals AI)Not measured51.2%Only Qwen3.8 Max (0902) measured
MedScribe (Vals AI)Not measured85.0%Only Qwen3.8 Max (0902) measured
Vals Multimodal IndexNot measured65.4%Only Qwen3.8 Max (0902) measured
IFBench72.5%Not measuredOnly Qwen3.5-35B-A3B measured
Long context reasoning (AA-LCR)72.0%80.3%Qwen3.8 Max (0902)
DTBench80.0%92.0%Qwen3.8 Max (0902)
LMCA29.5%46.2%Qwen3.8 Max (0902)
Furniture assemblyNot measured20.0%Only Qwen3.8 Max (0902) measured
LMArena Text13951483Qwen3.8 Max (0902)
LMArena Text: instruction following13791481Qwen3.8 Max (0902)
LMArena Text: creative writing13471477Qwen3.8 Max (0902)
LMArena VisionNot measured1313Only Qwen3.8 Max (0902) measured
LMArena Text: English14071491Qwen3.8 Max (0902)
LMArena Text: non-English13781473Qwen3.8 Max (0902)
LMArena Text: Chinese14571539Qwen3.8 Max (0902)
LMArena Text: German13671478Qwen3.8 Max (0902)
LMArena Text: Japanese13261473Qwen3.8 Max (0902)
LMArena Text: Korean13551460Qwen3.8 Max (0902)
LMArena Text: French14121498Qwen3.8 Max (0902)
LMArena Text: Spanish13901489Qwen3.8 Max (0902)
LMArena Text: Russian13761482Qwen3.8 Max (0902)
LMArena Text: Polish13851471Qwen3.8 Max (0902)
LMArena Text: multi-turn13901489Qwen3.8 Max (0902)
LMArena Text: longer queries13891488Qwen3.8 Max (0902)
LMArena Text: writing and language13651477Qwen3.8 Max (0902)
LMArena Text: entertainment and media13471464Qwen3.8 Max (0902)
LMArena Vision: EnglishNot measured1311Only Qwen3.8 Max (0902) measured
LMArena Vision: ChineseNot measured1358Only Qwen3.8 Max (0902) measured
LMArena Vision: reading textNot measured1322Only Qwen3.8 Max (0902) measured
LMArena Vision: humorNot measured1304Only Qwen3.8 Max (0902) measured
LMArena Vision: creative writingNot measured1312Only Qwen3.8 Max (0902) measured
Neither model has a result on the other 15 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Qwen3.5-35B-A3B and Qwen3.8 Max (0902)
BenchmarkQwen3.5-35B-A3BQwen3.8 Max (0902)Leader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

Qwen3.5-35B-A3B$0.0008Fits in one request
Qwen3.8 Max (0902)$0.0050Fits in one request

Qwen3.5-35B-A3B costs less

Repository review

50K input + 3K output tokens

Qwen3.5-35B-A3B$0.012Fits in one request
Qwen3.8 Max (0902)$0.118Fits in one request

Qwen3.5-35B-A3B costs less

Agent loop

200K input + 10K output tokens

Qwen3.5-35B-A3B$0.045Fits in one request
Qwen3.8 Max (0902)$0.460Fits in one request

Qwen3.5-35B-A3B costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of Qwen3.5-35B-A3B and Qwen3.8 Max (0902)
SpecificationQwen3.5-35B-A3BQwen3.8 Max (0902)
MakerQwenQwen
WeightsOpen weightsNo public weights listed
ReleasedFebruary 25, 2026September 3, 2026
Context window262K1M
Input $/Mtok$0.1625$2.00
Output $/Mtok$1.30$6.00
Coding indexNot benchmarked62.4
Independent measurements behind it15
Cost per solved taskNot computable$1.32
SourceOpenRouter listingOpenRouter listing

Questions

Is Qwen3.5-35B-A3B better than Qwen3.8 Max (0902) for coding?

Only Qwen3.8 Max (0902) has a coding score, 62.4 on the Vibe Coding Index; Qwen3.5-35B-A3B has none, so the two cannot be ranked on coding. Pick Qwen3.5-35B-A3B for output tokens 4.6x cheaper ($1.30 vs $6.00 per million).

Which is cheaper, Qwen3.5-35B-A3B or Qwen3.8 Max (0902)?

Qwen3.5-35B-A3B costs $0.1625 per million input tokens and $1.30 per million output tokens; Qwen3.8 Max (0902) costs $2.00 and $6.00. Cost per solved task is known for Qwen3.8 Max (0902) only.

Which has the bigger context window, Qwen3.5-35B-A3B or Qwen3.8 Max (0902)?

Qwen3.8 Max (0902), with 1M tokens against 262K.

How many benchmarks is the Qwen3.5-35B-A3B vs Qwen3.8 Max (0902) comparison based on?

41 benchmarks have a result for both models as of October 22, 2026. Qwen3.5-35B-A3B's coding index rests on 1 independent benchmark and Qwen3.8 Max (0902)'s on 5. Where only one model has a result, the page says so instead of naming a winner.

Compare next